There is an increasing call for population-wide environmental/policy interventions to increase physical activity despite the lack of large-scale intervention or epidemiological research documenting the benefits of such changes. This longitudinal study will link contemporaneous geographic locations of respondents with physical environment variables and data from an exceptional dataset including quality physical activity data. We will use four study years (1985, 1992, 1995, and 2001) of the Coronary Artery Risk Development in Young Adults Study [CARDIA], a longitudinal study of the antecedents and risk factors for cardiovascular disease in an ethnicity-, age- and sex-balanced cohort of 5,115 black and white young adults aged 18-30 years at baseline to examine relationships between environmental factors and physical activity. We will use complex longitudinal and spatial analytical models to explore relationships between environmental factors and physical activity. A critical element addressed will be residential self-selection, an issue of increasing concern as scholars attempt to understand how the environment affects physical activity. We will model physical activity as a function of covariates, some of which may be endogenous choices made by the individual. We will examine race/ethnic differentials in these effects and the impact of "the environment" shifts over time and through the lifecycle. The focus will be on examining how modifiable environmental factors will affect physical activity patterns among underserved communities and consequently will reduce ethnic and socioeconomic differentials in health status. The longitudinal analysis and the vast array of environmental measures used, coupled with the very high quality physical activity measures of CARDIA, allow us to capture the effects of the environment (and changes in location) on physical activity shifts. No study heretofore has had large-scale groupings and in-depth environmental measures over time to examine these issues in a dynamic manner.